Does Fibromuscular Dysplasia Cause Chronic Kidney Disease? A Two-Sample Mendelian Randomization Study

Overview

• Fibromuscular dysplasia (FMD) prevalence: ~ 3% (Rana and Al-Kindi 2021)
  
• Women are 85% to 90% of cases (Kim et al. 2013)
• Genome-wide association studies (GWAS) have identified variants that associate with FMD (Georges et al. 2021)
  
• Two-Sample Mendelian Randomization (MR) uses GWAS results to estimate a causal effect, subject to assumptions

FMD

• FMD affects artery walls
  
• Diagnosed with angiography, but may be overlooked in clinical practice
  
• Prevalence (U.S.A.): ~ 3%
  
• Women are 85% to 90% of cases

Chronic Kidney Disease (CKD)

• About one in ten people have CKD
  
• Affected ~ 16% of Americans aged 20+ in 1999 to 2004
  
• Causes and prevalences differ by ancestry
  
• Genome-wide association studies have revealed some genetic variants that affect CKD

FMD & CKD

• FMD is the second-most common abnormality of renal arteries (Twist, Leeuw, and Kroon 2018)
  
• FMD prevalence among potential kidney donors: 2% to 6%

Mendelian Randomization: Big Picture

• Major goal of biomedical research is to identify causes of diseases
  
• Randomized controlled trials are a gold standard for causal conclusions
  
• Confounding plagues observational studies that seek causal inferences
  
• Germline genotypes are randomly assigned at meiosis
  
• MR as “Nature’s randomized controlled trial”

Mendelian Randomization: A Method for Assessing Causal Relationships

Three Key Assumptions:

1. SNP instruments associate with Exposure
   
2. No common causes of SNPs and Outcome
   
3. No Exposure-independent path between SNPs and Outcome

• FMD is Exposure
  
• CKD is Outcome

Mendelian Randomization: A Method for Assessing Causal Relationships

Causal association between mTOR-dependent EIF-4E and EIF-4A circulating protein levels and type 2 diabetes: a Mendelian randomization study - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Mendelian-Randomization-unbiased-causal-associations-and-assumptions-Mendelian_fig5_344407603 [accessed 31 Mar, 2024]

FMD & CKD: MR Inputs

• Need GWAS Summary Statistics for FMD & CKD (with no subject overlap)
• FMD GWAS summary statistics (Georges et al. 2021)
  
• UK Biobank CKD GWAS summary statistics (women only) (Team 2018)
  
• Choose a set of SNPs based on associations with FMD (p-value threshold: \(10^{-8}\))

FMD & CKD: Results

FMD & CKD: Results

• Statistical power is presumably weak here

Summary & Next Steps

• No evidence of causal relationship between FMD and CKD (but statistical power!)
  
• Further sensitivity analyses are needed to quantify effects of possible biases
  
• Consider other sets of SNPs with different p-value thresholds

Acknowledgments

• Ganesh Laboratory, including Santhi Ganesh, Kristy Hunker, Min-Lee Yang, Trisha Gupte, and Sayhaan Goraya
• Xiang Zhou & Chang Xu
  
• Cardiovascular Medicine Training Grant (NIH T32-HL007853)

Thank you!

Fred Boehm: frederick.boehm@gmail.com

https://fboehm.us/

https://github.com/fboehm/2024-cgsi-ucla/